1. Symbolic Systems 100 Spring 2005
May 12, 2005: Conversation
IP Notice: Some slides from Julia Hirschberg and info from Cruse and
the readings
11/28/2018
Symbolic Systems 100 Spring 2005

2. Outline Conversation: a core phenomenon in high-level cognition
We’ve seen its role in computer science.
Today: Linguistics, Psychology, Philosophy
11/28/2018
SYMBSYS 100 Spring 2005

3. Turns and Utterances Conversation is characterized by turn-taking:
Speaker A talks, then speaker B, and so on.
How do speakers know:
Who should talk next
When they should talk
Evidence that language is structured to make this problem solvable: ‘split-second timing’
On overage, not much speaker overlap (~ 5% in English)
But little silence between turns either (less than 200 ms)
It takes at least this long to plan an utterance
So language must be set up in such a way that speakers can easily tell who should talk and when.
11/28/2018
SYMBSYS 100 Spring 2005

4. Turn Taking How do we know when a speaker is How do I know when
Giving up or taking a turn?
Holding the floor?
Interruptable?
How do I know when
Its my turn obligatorily
Optionally?
11/28/2018
SYMBSYS 100 Spring 2005

5. Simple Turn-Taking Rules
At each transition-relevance place (TRP) of each turn:
If current speaker has selected A as next speaker, then A must speak next
If current speaker does not select next speaker, any other speaker may take next turn
If no one else takes next turn, the current speaker may take next turn
TRPs are where the structure of the language allows speaker shifts to occur
11/28/2018
SYMBSYS 100 Spring 2005

6. Implications of the Turn- Taking Rules
Adjacency pairs set up next speaker expectations
GREETING/GREETING
QUESTION/ANSWER
COMPLIMENT/DOWNPLAYER
REQUEST/GRANT
11/28/2018
SYMBSYS 100 Spring 2005

7. Adjacency pairs: what to say?
Julia Hirschberg example:
Conversational partners expect certain patterns of behavior in normal conversation:
Julia: You got an A? That’s great!
Diane1: Yeah, I’m really smart you know.
Diane2: Well, I was just lucky I happened to read the chapter right before the test. Otherwise I never would have squeaked through.
11/28/2018
SYMBSYS 100 Spring 2005

8. Adjacency pairs: expected behavior
COMPLIMENTS are expected to be followed by DOWNPLAYERS.
It doesn’t have to happen
But deviation is significant.
11/28/2018
SYMBSYS 100 Spring 2005

9. More on deviation from expectations
Silence is different when it follows the first part of an adjacency pair
Significant silence is dispreferred
A: Is there something bothering you or not? (1.0s)
A: Yes or no? (1.5s)
A: Eh?
B: No.
11/28/2018
SYMBSYS 100 Spring 2005

10. Even more on deviation
We call a response that is not conversational appropriate or expected a “dispreferred” response.
Things like denying a request or saying no in general are dispreferred.
‘no’ to a simple request without explanation
Changing the topic abruptly without transition
Indicators:
“well”, “um”, silence
11/28/2018
SYMBSYS 100 Spring 2005

11. Grounding and Contributions
Dialogue is a collective act performed by speaker and hearer
Common ground: set of things mutually believed by both speaker and hearer
Need to achieve common ground, so hearer must ground or acknowledge speakers utterance.
Clark (1996):
Principle of closure. Agents performing an action require evidence, sufficient for current purposes, that they have succeeded in performing it
(Interestingly, Clark points out that this idea draws from Norman (1988) work on non-linguistic acts)
Need to know whether an action succeeded or failed
11/28/2018
SYMBSYS 100 Spring 2005

12. How is Closure Achieved?
Clark and Schaefer
Each joint linguistic act is a “contribution”
Two parts:
Presentation
acceptance
11/28/2018
SYMBSYS 100 Spring 2005

13. 11/28/2018
SYMBSYS 100 Spring 2005

14. Clark and Schaefer: Grounding
Continued attention: B continues attending to A
Relevant next contribution: B starts in on next relevant contribution
Acknowledgement: B nods or says continuer like uh-huh, yeah, assessment (great!)
Demonstration: B demonstrates understanding A by paraphrasing or reformulating A’s contribution, or by collaboratively completing A’s utterance
Display: B displays verbatim all or part of A’s presentation
11/28/2018
SYMBSYS 100 Spring 2005

15. A human-human conversation
11/28/2018
SYMBSYS 100 Spring 2005

16. Grounding examples Display: Acknowledgement C: I need to travel in May
A: And, what day in May did you want to travel?
Acknowledgement
C: He wants to fly from Boston
A: mm-hmm
C: to Baltimore Washington International
[Mm-hmm (usually transcribed “uh-huh”) is a backchannel, continuer, or acknowledgement token]
11/28/2018
SYMBSYS 100 Spring 2005

17. Grounding Examples (2) Next relevant contribution:
A: And you’re flying into what city?
Q: Seattle
The second pair part of an adjacency pair grounds the first pair
11/28/2018
SYMBSYS 100 Spring 2005

18. Grounding Examples (3) Acknowledgement + next relevant contribution
And, what day in May did you want to travel?
And you’re flying into what city?
And what time would you like to leave?
The and indicates to the client that agent has successfully understood answer to the last question.
11/28/2018
SYMBSYS 100 Spring 2005

19. Admin break *** Symbolic Systems Forum ***
Mind Out of Matter: A History of the Quest for a Conscious Machine
Jessica Riskin, History Department
Thursday, May 12th at 4:15pm
Room C
Stanford University
11/28/2018
SYMBSYS 100 Spring 2005

20. Admin break: Riskin Abstract
The talk will describe my book-in-progress, which is about the genesis and early history of artificial life. The book will examine attempts to simulate the behaviors and bodily functions of living creatures from the first appearance of these attempts as thought-experiments in the mid-seventeenth century, through their transformation into actual experiments in the early eighteenth century, and ending with the widespread rejection, in the early nineteenth century, of the possibility of simulating life in mechanism. My central interest in telling this story is in the continual re-definition of life and mind, on the one hand, and of machinery, on the other, resulting from these measurements of each against the other. The terms in which nineteenth-century philosophers and engineers rejected the possibility of mechanical simulations of life implied a new understanding of life's (and machinery's) defining features. Some key elements of this new understanding would later, I believe, inform the return of artificial life in cybernetics.
11/28/2018
SYMBSYS 100 Spring 2005

21. Admin Break: Homework Original policy: late homework not accepted
You guys have asked to accept late homework with penalty
OK, seems reasonable. New policy:
Late homework will be accepted for 24 hours after due date
Remember that due date is 10:00am.
Not 11:30am.
But late homework will have a 40% penalty.
11/28/2018
SYMBSYS 100 Spring 2005

22. Conversation structure & expectations depend on task
Telephone Openings
Pat: Hell?
Chris: Hi, Pat. It’s Chris.
Pat: Hi!
Telephone closings (6-turn)
Chris: Well, I just wanted to see how you were doing
Pat: Thanks for calling. We’ll have to have lunch sometime
Chris. I’d like to
Pat: OK
Chris: OK
Pat: See you
Chris: Yeah, see you
11/28/2018
SYMBSYS 100 Spring 2005
From Julia Hirschberg

23. Conversation structure & expectations depend on task
Service Encounters
Clerk: Good morning. Is there something I can help you with?
Pat: Hi. Yeah. I wonder if you could show me
Meetings
Boss: Today I want to focus on next year’s goal statements Chris, could you report please…
Chris:…
Boss: Pat, now let’s hear from yhou..
Pat:…
What about?
IM? ?
11/28/2018
SYMBSYS 100 Spring 2005
From Julia Hirschberg

24. Presequences
11/28/2018
SYMBSYS 100 Spring 2005

25. Pre-requests
A: Hi. Do you have uh size C flashlight batteries [pre-request]
B: Yes sir [go ahead]
A: I’ll have four please [request]
B: ((turns to get)) [response]
11/28/2018
SYMBSYS 100 Spring 2005

26. Why presequences Avoid telling someone something they know already
Put an announcement in a focused position
Avoid a dispreferred response
11/28/2018
SYMBSYS 100 Spring 2005

27. Disfluencies
11/28/2018
SYMBSYS 100 Spring 2005

28. Disfluencies: standard terminology (Levelt)
Reparandum: thing repaired
Interruption point (IP): where speaker breaks off
Editing phase (edit terms): uh, I mean, you know
Repair: fluent continuation
11/28/2018
SYMBSYS 100 Spring 2005

29. Kinds of disfluencies Fillers (uh, um) Edit terms (I mean, you know)
Repetitions
Pauses
11/28/2018
SYMBSYS 100 Spring 2005

30. Counts (from Shriberg, Heeman)
Sentence disfluency rate
Levelt human dialogs: 34% of sentences disfluent
Switchboard telephone convesrations: ~50% of multiword sentences disfluent
Word disfluency rate
Switchboard telephone chats: 6%
Booking air travel: %
11/28/2018
SYMBSYS 100 Spring 2005

31. Disfluencies Clark and Fox Tree Looked at “um” and “uh”
“uh” includes “er” (“er” is just British non-rhotic dialect spelling for “uh”)
Different meanings
Uh: used to announce minor delays
Preceded and followed by shorter pauses
Um: used to announce major delays
Preceded and followed by longer pauses
11/28/2018
SYMBSYS 100 Spring 2005

32. Um versus uh: delays (Clark and Fox Tree)
11/28/2018
SYMBSYS 100 Spring 2005

33. Utterance Planning
The more difficulty speakers have in planning, the more delays
Consider 3 locations:
I: before intonation phrase: hardest
II: after first word of intonation phrase: easier
III: later: easiest
And then uh somebody said, . [I] but um -- [II] don’t you think there’s evidence of this, in the twelfth - [III] and thirteenth centuries?
11/28/2018
SYMBSYS 100 Spring 2005

34. Delays at different points in phrase
11/28/2018
SYMBSYS 100 Spring 2005

35. Backchannels Uh-huh Yeah What are the differences between these?
11/28/2018
SYMBSYS 100 Spring 2005

36. Conversational Implicature
A: And, what day in May did you want to travel?
C: OK, uh, I need to be there for a meeting that’s from the 12th to the 15th.
Note that client did not answer question.
Meaning of client’s sentence:
Meeting
Start-of-meeting: 12th
End-of-meeting: 15th
Doesn’t say anything about flying!!!!!
What is it that licenses agent to infer that client is mentioning this meeting so as to inform the agent of the travel dates?
11/28/2018
SYMBSYS 100 Spring 2005

37. Conversational Implicature (2)
A: … there’s 3 non-stops today.
This would still be true if 7 non-stops today.
But no, the agent means: 3 and only 3.
How can client infer that agent means:
only 3
11/28/2018
SYMBSYS 100 Spring 2005

38. Grice: conversational implicature
Implicature means a particular class of licensed inferences.
Grice (1975) proposed that what enables hearers to draw correct inferences is:
Cooperative Principle
This is a tacit agreement by speakers and listeners to cooperate in communication
11/28/2018
SYMBSYS 100 Spring 2005

39. 4 Gricean Maxims Relevance: Be relevant
Quantity: Do not make your contribution less (1) or more (2) informative than required
Quality: try to make your contribution one that is true (don’t say things that are false or for which you lack adequate evidence)
Manner: Avoid ambiguity and obscurity; be brief and orderly
11/28/2018
SYMBSYS 100 Spring 2005

40. Relevance A: Is Regina here? B: Her car is outside. Implication: yes
Hearer thinks: why would he mention the car? It must be relevant. How could it be relevant? It could since if her car is here she is probably here.
Client: I need to be there for a meeting that’s from the 12th to the 15th
Hearer thinks: Speaker is following maxims, would only have mentioned meeting if it was relevant. How could meeting be relevant? If client meant me to understand that he had to depart in time for the mtg.
11/28/2018
SYMBSYS 100 Spring 2005

41. Quantity B: I have 5 dollars
A:How much money do you have on you?
B: I have 5 dollars
Implication: not 6 dollars
Similarly, 3 non stops can’t mean 7 non-stops (hearer thinks:
if speaker meant 7 non-stops she would have said 7 non-stops
A: Did you do the reading for today’s class?
B: I intended to
Implication: No
B’s answer would be true if B intended to do the reading AND did the reading, but would then violate maxim
11/28/2018
SYMBSYS 100 Spring 2005

42. Quantity What did you have for lunch today? Baked beans on toast Food
Violates quantity 1
87 warmed-up baked beans served on a slice of toast 12.7x10.3cm, which had been unevenly toasted
Violates quantity 2
11/28/2018
SYMBSYS 100 Spring 2005

43. Manner
Avoid obscurity, ambiguity, unnecessary prolixity and be orderly.
Miss X produced a series of sounds that corresponded closely with the score of “Home Sweet Home”
11/28/2018
SYMBSYS 100 Spring 2005

44. What can you do with Maxims
Obey them
Violate them (I.e. deceive)
Face a conflict (I’d like to tell you more, but I don’t have the information)
Exploit them: I.e. intentionally violate them to signal a hidden meaning:
http/
11/28/2018
SYMBSYS 100 Spring 2005

45. How to deceive without lying
Mom: Where are you going?
Little George: I’m going to the candy store.
If little George is also going to the pool hall, his sentence is true.
But what it implicates is false
This implicates (via QUANTITY) something that is false, I.e. that I am ONLY going to the candy store
11/28/2018
SYMBSYS 100 Spring 2005

46. What is said versus implicatures
What is said + what is implicated = overall meaning
“What is said” = proposition
Propositions can be challenged or agreed with:
A: It’s raining
B: That’s not true
A: The correct answer is ‘7’
B: I agree
11/28/2018
SYMBSYS 100 Spring 2005
Text from web from lecture on Cruse

47. What is said vs. implicatures
Implicatures (symbolized with +>) cannot be directly challenged or agreed with:
A: Theo is an excellent typist and has a sunny disposition
+> Theo is not a very good cognitive scientist.
B: “That’s not true./I agree”
Not: B denies/agrees that Theo is not a very good linguist
11/28/2018
SYMBSYS 100 Spring 2005

48. Conclusions Conversation
A really key phenomenon of high-level cognition
Studied in linguistics, philosophy, psychology, computer science
Study is still only in its infancy!
I hope some of you will be inspired to look into conversation!
11/28/2018
SYMBSYS 100 Spring 2005